Introduction
The Word Hub Limited leverages artificial intelligence (AI) technology, where appropriate, to enhance our translation services while upholding the highest standards of quality, privacy, and ethics. This AI Policy Statement outlines our practices and guidelines for integrating and using AI tools in our translation processes.
Purpose
The purpose of this policy is to:
- ensure the responsible and ethical use of AI in translation.
- protect client data and maintain privacy.
- provide transparency about our AI practices.
- maintain high standards of translation accuracy and quality.
Scope
This policy applies to all employees, linguists, and third-party vendors involved in the development, implementation, and use of AI technologies within The Word Hub Limited.
- Data Privacy and Security
1.1 AI Vendor
Our chosen third-party vendor of AI output for use in our translation processes is an ISO 27001:2013 certified company. Any data inputted into the system is protected using state-of-the-art encryption. In addition, translations carried out by The Word Hub Limited using this service are not used to train baseline translation engines or other clients’ engines.
1.2 Translation Memories
Any translation memories used by The Word Hub Limited to support the AI output generated are client-specific and not used to benefit the work of other clients.
1.3 Sensitive Data
The Word Hub Limited should be made aware of any data of a sensitive nature that should not be processed using AI upon receipt of the initial quotation request.
1.4 Cybersecurity
All computers and software used by The Word Hub Limited are protected by firewalls and anti-virus software, as well as updated to the latest versions.
- Quality Assurance 2.1 AI In The Loop
- All AI translations generated are reviewed and edited by The Word Hub Limited’s trusted professional linguists to ensure accuracy and quality.
2.2 Baseline Engines
- Baseline engines are trained using billions of words of public data and public corpora to enhance translation quality.
- Regular testing of baseline engines is carried out to evaluate quality. Human linguists are used for these evaluations; automatic evaluation methods are not used.
- Our chosen provider uses Neural Machine Translation (NMT). In contrast to Large Language Models (LLMs) which are trained on vast quantities of monolingual data to predict text, NMT engines are trained using bilingual parallel corpora to learn and produce text.
2.3 Client feedback
- Clients are encouraged to provide feedback on AI-generated translations. This feedback will be used to improve our AI tools.
2.4 Internal workflows and audits
- The Word Hub Limited develops and implements robust internal workflows to ensure consistent and accurate AI implementation.
- Workflows are regularly audited and attested by The Word Hub Limited’s ISO 9001:2015 and ISO 17100:2015 certifications.
- Ethical Use of AI
3.1 Bias Mitigation
- We actively work to identify and reduce biases in our AI translation models through regular audits in accordance with our ISO 9001:2015 and ISO 17100:2015
certifications.
3.2 Transparency
- Clients will be informed when AI tools are used in the translation process. • Clients have the right to make a Freedom of Information request for more details about our use of AI technology.
3.3 Accountability
- Human translators are always responsible for the final output, even when AI tools are utilised.
- Client Consent 4.1 Opt-Out Option
- Clients have the option to opt out of having their data used for AI training, if desired. 5. Employee Training and Development
5.1 Training
- All of our trusted professional linguists are experienced in the use of AI tools and the editing of their output.
- Where suitable, The Word Hub Limited employees take part in training programs and join focus groups about AI advancements and best practices in translation.
5.2 Continual Learning
- All The Word Hub Limited employees keep up-to-date with the latest news relating to the use of AI in translation.
- Intellectual Property
6.1 Ownership
- Despite using the services of a third-party AI vendor that is available in the public domain, ownership of AI-generated translations and associated intellectual property is clearly defined as belonging to The Word Hub.
6.2 Usage Rights
- The usage and sharing of AI tools and outputs are specified and managed appropriately.
- Vendor Management
7.1 Vendor Evaluation
- Third-party AI vendors are assessed for security, reliability, and ethical practices. 7.2 Contractual Obligations
- Contracts with AI vendors include clauses related to data privacy, security, and ethical use.
- Policy Review and Updates
This policy will be reviewed annually and updated as necessary to reflect changes in technology, regulations, and company practices.
- Contact Information
For any questions or concerns regarding this AI Policy Statement, please contact:
- Name: Darren Verdi-Ashton
- • Email: [email protected]
